Metabolic profiles of the volatile constituents of body fluids, urinary acids, and plasma and urinary steroids will be determined for selected disease states by high-precision GC. The automated system will be used for reproducible recording of profiles on a repetitive basis, and the data acquisition. Data obtained from sufficient numbers of "normal" and pathological samples will be further analyzed by large computers to find out whether metabolic patterns are characteristic for certain conditions. Pattern recognition methods that will be investigated include the use of binary classifiers and cluster analysis. Necessary identifications of profile constituents will be made by means of capillary GC-MS. Glass capillary column techniques will further be used for development of additional profiling procedures for acidic basic volatile body fluid constituents. Simplification of profiles will be approached through the use of subtractive columns, column-switching techniques, and selective detectors. Both headspace samples and extracts of body fluids will be studied. These studies are likely to contribute to a wider use of GC methods for biomedical studies as well as better understanding of altered metabolic conditions in disease states.